Monitoring active fluvial sand and gravel mining operations in the lower Mekong Basin using animated time-series of Sentinel-2 imagery

Abstract

The Mekong Basin is a global centre for sand and gravel extraction and the intensity of such mining is already contributing towards severe environmental degradation. However, the region lacks effective infrastructure for monitoring the industry. Thus, a novel exploratory interactive time-series animation approach was trialled at five study sites within the Basin, in an attempt to overcome longstanding limitations of Earth observation-based methods for fluvial sand and gravel detection and monitoring. Animation was deployed in an attempt to leverage the relatively high temporal-resolution of Sentinel-2-MSI RGB imagery, as a means of augmenting the informational power of its spatial-resolution, in the hope of identifying characteristic spatio-temporal traces of small-scale active fluvial aggregate mining features that are indistinguishable in its spatial-resolution alone. The method was implemented entirely within the Google Earth Engine Code Editor, by means of a 3rd-party interactive slideshow animation package. Distinguishable features of fluvial aggregate mining at this spatio-temporal resolution included: mining dredger/barge clusters; pipelines for transporting aggregate from dredgers to the shore; artificial bar incision and bank indent; mining operational bases; sediment storage/drying piles; excavators; and near-constant small-scale topographic changes that produce a ‘flickering’ effect. Some of these features are too-small to easily distinguish within static frames, but become distinguishable when their temporal-dimension is visualised through animation. When these features are compiled into a suit of indicators of extraction and post-extraction processing, one can use identified collocation of indicators of both types to gain a sense of whether a suspected mining site is in active operation.